The present invention relates to processing video frames, and more particularly, to a brightness adjusting methods for video frames of a video sequence and the related devices thereof.
While processing video sequences, in which the video sequence comprises a series of video frames, it is possible to collect pixels' statistics corresponding to each video frame, for example, the luminance value (Y) and chrominance values (U, V) of each pixel. Based on the collected statistics, it is possible to manipulate the distributions to make the video picture more favorable to users. For example, normalizing the luminance distribution to expand the range of distribution, and thus enhances the contrast of the video frame. This process of manipulating the display characteristics is well known to those of average skill in the art and therefore additional explanatory details are omitted for the sake of brevity. Many methods are well documented in the related art. Any such method is equally valid as an example herein.
It is well known to those of average skill in this art that when many video frames are similar but still have different brightness distributions, said conventional normalization process for manipulating the brightness distribution of the entire picture could cause some points to blink. The blinking effect is visible to a viewer of the series of video frames.
The related art utilizes noise suppressing techniques such as an IIR filter (Infinite Impulse Response filter) to help reduce the blinking problem. This is commonly called smoothing. However, there are problems with this related art approach. First, while the blinking is eliminated, unwanted smoothing of the scene transitions (i.e., scene changes) is introduced. The smoothing operation of the related art is not selectively applied and as a result, the conventional methods and techniques incorrectly smoothes content change of video frames in the video sequence when those video frames are not similar but different from each other.